| Acknowledge (ack) | Marking an open issue of a scheduled job as "checked" to close it. An administrator does it, because it requires running a command. The job keeps running even if no one acknowledges it. | 6.1. Scheduled Jobs |
| Administrator | A person who manages users, models, role templates, and notices. Some screens are visible only to administrators. This person may be different from the operator, who installs and runs the server. | Administrator Manual |
| Allowed user | A person who is allowed to talk to the bot. You set this with the user ID used in the messenger. | 6.2. Messenger Connections |
| Approval request | A request for permission before an AI agent does something that may be risky, such as deleting a file or running a program. The AI agent waits until you answer. | 2.6. Answering Approvals and Questions |
| Bot | An account that speaks for your AI agent inside a messenger. You create one separately for each messenger. One bot is connected to only one profile. | 6.2. Messenger Connections |
| Bot token | A secret string that a messenger issues for each bot. AAS does not show the value again after you enter it. | 6.2. Messenger Connections |
| Change log | A record of who changed what and when. Secret values and text contents are not recorded. | 7.2. Change Log |
| Channel | One messenger connected to a profile. On the screen, "channel" and "messenger" mean the same thing. | 6.2. Messenger Connections |
| Clarifying question | A question that an AI agent asks the user in order to continue its work. On the screen it appears as a "Question from the AI Agent" card. | 2.6. Answering Approvals and Questions |
| Connection scope | A set of permissions you choose when connecting Claude Code to AAS. The choices are "Read only", "Skill development", and "Everything (my AI agents)". You cannot change it after connecting. | 7.5. Connecting Claude Code |
| Dashboard | The first screen that gathers the status of your AI agent workspace and the items that need action on one screen. | 1.1. Introducing AAS and Signing In |
| Department | The organization unit a user belongs to. An administrator sets it. Each department has one banner text in the 3D Office. | 7.6. Company Information and My Account |
| Drawer | A detail window that opens on the right side of the screen when you click a row in a list. This manual uses this name. | 1.2. Using Common Screen Features |
| Gateway | The channel through which an AI agent receives messenger messages and scheduled jobs. The gateway must be running for messengers and scheduled jobs to work. An administrator starts and stops it. | 6.2. Messenger Connections |
| Instructions (AGENTS) | Text that describes how an AI agent should work. It covers the way of working, the report format, and what not to do. It applies from a new chat. | 4.3. Personality (SOUL) and Instructions (AGENTS) |
| Issue | A record that AAS keeps when a scheduled job fails. It closes only after a person acknowledges (ack) it. | 6.1. Scheduled Jobs |
| Kanban | A board where you turn work into cards, place them in columns, and assign them to AI agents. | 3.3. Handing Off Work with Kanban |
| Main profile | The mark on the profile (main) that each workspace has at the start. It cannot be renamed or deleted. | 4.1. AI Agent Workspaces |
| Meeting room | A place where people and several AI agents work together in one chat room. | 3.2. Meeting Rooms and TFT Rooms |
| Messenger connection | Connecting a messenger such as Telegram, Discord, Slack, or Mattermost so that you can talk to your AI agent there. It is also the name of a screen. | 6.2. Messenger Connections |
| Model | The AI that an AI agent uses to read and write text. Each profile has a main model. | 4.2. Profile Overview and Model |
| Notice | A message that an administrator posts in the bar at the top of the screen to announce maintenance, outages, or new features. | 7.6. Company Information and My Account |
| Office | A 3D screen where AI agents and people work together in one room. | 3.1. Touring the 3D Office |
| Operator | A person who installs AAS and manages the server. This person may be different from the administrator who manages users on the screen. An administrator hands server-side work that cannot be done on the screen to the operator. | Administrator Manual |
| Pairing | An approval request that appears when a person who is not allowed talks to the bot for the first time, or the process of approving it. | 6.2. Messenger Connections |
| Personality (SOUL) | The introduction that an AI agent always follows. It describes its role, tone of voice, and the rules to keep. | 4.3. Personality (SOUL) and Instructions (AGENTS) |
| Profile | The set of settings for one AI agent in a workspace. It holds the personality, model, skills, tools, scheduled jobs, and messenger connections. Each workspace has a main profile (main) at the start. | 4.2. Profile Overview and Model |
| Role template | A personality (SOUL) form written in advance for each company role. | 4.3. Personality (SOUL) and Instructions (AGENTS) |
| Safeguard | A feature that blocks risky commands. A blocked command is shown in the chat as "blocked by a safeguard". | 2.7. When Replies Are Blocked or Slow |
| Scheduled job (cron) | A feature that makes an AI agent do the same task automatically at a set time. | 6.1. Scheduled Jobs |
| Secret | An API key, token, or password that an AI agent uses to connect to an outside service. Once entered, it is not shown on the screen again. | 4.4. Secrets and Credential Files |
| Session | One chat. Chats, scheduled jobs, and messenger conversations each create sessions. | 7.1. Usage |
| Skill | A bundle that describes how an AI agent does a particular task. Once installed, the AI agent can do that task. | 5.1. Skill List |
| Skill Hub | The place where you find and install skill packages and publish your own skills. | 5.3. Installing from the Skill Hub |
| Skill map | A screen that shows in a diagram where a skill is called from and what it calls. | 5.6. Skill Map |
| Team manager | A person who can give work to AI agents and change settings in the AI agent workspaces of teammates in the same department. Changes are recorded in the owner's log with the manager's name. | 7.2. Change Log |
| Token (access) | A secret string that carries access rights. There are tokens for Claude Code connections and tokens for outside services. It is different from "Token (AI)". | 7.4. API Keys and Tokens |
| Token (AI) | A unit that counts how much text the AI read and wrote. The amount read is input tokens, and the amount written is output tokens. | 7.1. Usage |
| Tool | A feature an AI agent uses to do its work, such as reading a file, running a command, or calling an outside service. | 4.5. Connecting External Tools (MCP) |
| Toolset | Tools of the same kind grouped together. You can turn each toolset on or off, and when it is off the AI agent does not use the tools in that set. | 4.5. Connecting External Tools (MCP) |
| Usage | A screen that shows the number of chat sessions, messages, and tokens. | 7.1. Usage |
| Work folder | The place where files you upload and files that AI agents create are collected inside an AI agent workspace. Profiles in the same workspace share it. | 2.3. Uploading and Downloading Files |
| Workspace owner | The user who owns an AI agent workspace. The owner can change the workspace's settings. | 4.1. AI Agent Workspaces |